Euclidean Technologies

Euclidean Technologies put sequence-to-sequence language models, the architecture class behind modern chatbots, to work on stock selection back in 2020, well ahead of the generative-AI wave. John Alberg and Michael Seckler founded the Seattle firm in 2008 around machine learning for long-horizon, deep-value equity investing.

Its models do two jobs: estimate intrinsic value to surface undervalued U.S. companies, and flag cheap-looking stocks likely to keep underperforming so the portfolio can avoid them. In 2023 the firm converted its private funds into the Euclidean Fundamental Value ETF on NYSE Arca, and it continues to research language models for reading SEC filings, earnings transcripts, and analyst reports.